LLM Stats vs Vellum: What Each Does and What It Costs

Q: Does Model Odds send email alerts?

Yes. You can sign up with a confirmed email to get alerts . Every subscriber gets every company's releases at the same time . Each Anthropic and OpenAI company also has its own page that lists every past release , which you can check alongside the alerts.

Compare LLM Stats and Vellum on what each does, what each costs and what to check before you pick one.

By SupaKeeper · 2026-10-30

SupaKeeper is published by LunarWerx, which makes some of the tools compared here.

LLM Stats combines benchmark results and prices into one ranking of models ^[1], while Vellum is an assistant that asks for explicit approval before high-stakes actions ^[2].

Option	What its sources say it does	Cost	Account or install
LLM Stats	Aggregates benchmark results, such as GPQA and SWE-Bench Verified, with coding-arena performance and pricing into one ranking ^[1]. Tracks 408 canonical models ^[1].	not published	not published
Vellum	The assistant can send emails, manage calendars, browse the web, create slides, write code, run deep research, monitor websites and run recurring tasks in the background ^[2]. By default it runs locally ^[2].	Free to download and start ^[2]. New users get their first tasks free ^[3]. Vellum Cloud subscriptions start at $30 per month for Mighty, $100 per month for Super or $200 per month for Ultra, with dedicated compute, storage and bundled credits ^[2].	Available on Mac, iOS, Android, web app, voice, email, Telegram, Slack and terminal ^[2]. Open source, so its code can be inspected ^[2].
Model Odds	Ranks every Anthropic and OpenAI model family by its estimated chance of a release within 30 days ^[4]. Each company has its own page that lists every past release ^[4].	Free, and so is its API, with no account needed ^[4].	A web app that runs in the browser and needs JavaScript ^[4]. It is a website with a public API ^[4].

The short answer

LLM Stats aggregates benchmark results such as GPQA and SWE-Bench Verified, coding-arena performance and pricing into one comparable ranking of models ^[1]. It pulls each provider's price from its public price list and checks it against billing samples ^[1]. Vellum is an assistant, and it asks for explicit approval before high-stakes actions, such as sending emails to new contacts or spending credits, until you mark them as trusted ^[2]. By default the assistant runs locally ^[2]. If you need model rankings and prices in one place, the LLM Stats facts cover that ^[1]. If you need an assistant that runs locally by default and asks before risky actions, the Vellum facts cover that ^[2] ^[2].

What LLM Stats and Vellum are built for

LLM Stats offers a weekly curated digest of models, benchmarks and analysis, delivered to your inbox once a week ^[1]. Its score combines benchmark results and coding-arena performance with pricing into one comparable ranking ^[1]. Its prices come from each provider's public price list and are checked against billing samples ^[1]. By default the Vellum assistant runs locally on the user's machine, and Vellum Cloud runs each assistant in an isolated container provisioned for that user ^[2]. Vellum asks for explicit approval before high-stakes actions run, such as sending emails to new contacts or spending credits, until the user marks them as trusted ^[2].

What Model Odds is built for

Model Odds is made by the team that publishes this site. It ranks every Anthropic and OpenAI model family by its estimated chance of a release within 30 days ^[4]. Each company has its own page that lists every past release ^[4]. Each past launch it uses has to be confirmed by an official post and an independent timeline, and each is listed with its date ^[4]. It runs three estimators side by side and publishes the spread between them as the error bar ^[4]. The estimate is built on the gaps between recent launches and takes into account how long the current wait has already lasted ^[4]. Signals read automatically from company SDKs and docs, and from third-party model catalogues when two of them agree, can only raise the odds ^[4]. The estimator can replay any past date without using data from after that date ^[4]. You can sign up with a confirmed email to get alerts, and every subscriber gets every company's releases at the same time ^[4].

Pricing and fees

Model Odds is free, and so is its API, with no account needed ^[4]. The facts for LLM Stats do not state a fee for its own use, so its cost is not published here, and its own pages are the place to check it. Vellum's plan prices are set out in the table above, with each price tied to its own plan.

Setting it up and daily use

Model Odds is a web app that runs in the browser and needs JavaScript ^[4]. Its public API returns the odds for each model line, the launch record and service health ^[4]. You can sign up with a confirmed email to get alerts ^[4]. For LLM Stats, the facts describe a weekly newsletter digest delivered to your inbox ^[1]. Its prices are pulled from provider price lists ^[1]. The facts here give no setup steps for LLM Stats, so check its pages for those. By default Vellum runs locally, and Vellum Cloud runs each assistant in an isolated container ^[2]. Before high-stakes actions run, Vellum asks for explicit approval until you mark them as trusted ^[2]. The facts here give no setup steps for Vellum either, so check its pages for those.

Things to check before you choose

Ask what you need the tool to answer. Model Odds estimates the chance of a release within 30 days for each Anthropic and OpenAI model family ^[4], while LLM Stats combines benchmark results and prices into one score ^[1]. Check how current the figures are. The Model Odds website and API share one snapshot that is cached for up to five minutes ^[4]. The LLM Stats score is built from benchmark results, so check the date of the results you rely on ^[1]. Check what costs money. Model Odds is free ^[4], the facts here give no price for LLM Stats, so check its own pages, and Vellum's plan prices are in the table above. Check how approvals work. Vellum asks for explicit approval before high-stakes actions, until you mark them as trusted ^[2], and by default it runs locally ^[2].

How to choose

If you need to know when a model family is likely to release, Model Odds is built for that ^[4]. If you need one ranking of models by score and price, LLM Stats is built around that ^[1]. If you need an assistant that runs locally by default and asks before risky actions, Vellum is built around that ^[2] ^[2]. Each of these answers a different question, so start with the one you need answered.

Frequently asked questions

Is Model Odds free to use?

Yes. Model Odds is free, and so is its API, with no account needed ^[4]. It is a web app that runs in the browser and needs JavaScript ^[4]. Its public API returns the odds for each model line, the launch record and service health ^[4]. The website and the API share one snapshot that is cached for up to five minutes, so the figures you read can be up to five minutes old ^[4].

What does Model Odds estimate for each model family?

Model Odds ranks every Anthropic and OpenAI model family by its estimated chance of a release within 30 days ^[4]. It runs three estimators side by side and publishes the spread between them as the error bar ^[4]. The estimate is built on the gaps between recent launches and takes into account how long the current wait has already lasted ^[4].

Does Vellum run its assistant locally by default?

Yes. By default the assistant runs locally, and Vellum Cloud is the option where each assistant runs in an isolated container provisioned for that user ^[2]. Before high-stakes actions run, such as sending emails to new contacts or spending credits, Vellum asks for explicit approval until the user marks them as trusted ^[2]. Check Vellum's own pages for details beyond that.

How does LLM Stats get its prices?

LLM Stats pulls pricing from each provider's public price list and cross-checks it against billing samples through the LLM Stats proxy ^[1]. Its score is a separate feature that aggregates benchmark results such as GPQA and SWE-Bench Verified, coding-arena performance and pricing into one comparable ranking ^[1]. The facts here do not give a price for LLM Stats itself, so check its own pages for that.

Does Model Odds send email alerts?

Yes. You can sign up with a confirmed email to get alerts ^[4]. Every subscriber gets every company's releases at the same time ^[4]. Each Anthropic and OpenAI company also has its own page that lists every past release ^[4], which you can check alongside the alerts.

Model Odds is free, and so is its API, with no account needed. ^[4]

Try Model Odds